CASE STUDY - An Alternate Perspective on Extract, Transform, and Load Improves Business Intelligence
By Thomas Hooker, Product Marketing Manager, Cisco
The extract, transform, and load (ETL) process that populates a data warehouse offers ample room for efficiency. Tapping this well of efficiency, however, requires a small but powerful shift in perspective on how to schedule the jobs that comprise ETL processes.
Consider the fact that ETL processes are often scheduled using job schedulers that do not easily extend to jobs beyond the data integration platform. Yet, ironically, ETL processes must extend, because BI requires data from diverse applications.
Shifting from a focused point-solution model and managing ETL processes from a broad enterprise automation perspective would obviously yield smoother ETL workflows across diverse systems. Less effort, greater results—a perennial IT intent. Many BI teams are discovering that when they shift their perspective on ETL job scheduling to an enterprise-oriented view, they tap into efficiency, along with greater control and visibility.
A case in point: you may have heard of Electronic Arts (EA), a global leader in interactive entertainment. Did you know that timely, accurate business intelligence (BI) about products, sales, customers, and finance helps the company stay on top of its game?
For example, EA sells products both directly and through retail outlets, such as Walmart, and uses good BI to understand which games are selling well in each of its markets. That kind of BI requires massive, detailed data to be processed quickly. The company’s BI needs are growing, so the efficiency of the underlying ETL processes needed to increase to process more information in the existing time window.
The company relies on Informatica PowerCenter to integrate its data. A lot of data required to populate the data warehouse resides with varied enterprise applications. Personnel used to monitor all the broad-reaching, interdependent jobs to keep processes flowing in the correct sequence and prevent jobs from halting mid-process and causing delays. To become more efficient, they needed a script-free way to schedule Informatica jobs along with jobs that cross other technologies.
The team took a broad perspective and created one enterprise view and control point over scheduling all the ETL processes that populate the data warehouse, with automated monitoring and alerting for all the processes—an automated enterprise scheduling model.
Shift in Perspective, Shift in Results
The ability to quickly schedule jobs for any application and platform was the key to improving ETL efficiency. The company deployed the enterprise scheduler from Cisco because the software handles jobs across the data center and offers many connection options, including Informatica and Oracle connections. Users visually create job workflows that span applications and systems without having to write scripts.
With the enterprise job scheduling model in place, ETL job flow improved right away. This broader visibility and automation saves the team many hours per day, and ETL jobs run faster and more efficiently because lag times between processes are reduced or eliminated. Potential errors are flagged. Important data and reports for finance, sales, and planning make it to their destination on time. With more efficient ETL, IT personnel are free to better serve users and support strategic business objectives.
The overarching results of automating ETL processes with an enterprise job scheduling approach are significant time savings and greater control of BI processes, which helps the team scale BI with confidence and enables the business to keep its market responsiveness sharp.
To read a detailed case study about this implementation or to learn more about how to improve ETL using Cisco® Tidal Enterprise Scheduler, please visit www.cisco.com.
This article originally appeared in the issue of .